analysis/capt_plot_nyt_amazon_comparison.py

from __future__ import annotations

import argparse
import csv
import html
import re
from dataclasses import dataclass
from pathlib import Path
from typing import Dict, List, Optional, Sequence, Set, Tuple

from capt_validation_schema import ensure_directory


NOT_ON_LIST_RANK = 16
CHART_WIDTH = 1120
CHART_HEIGHT = 420
CHART_LEFT = 68
CHART_TOP = 28
CHART_RIGHT = 24
CHART_BOTTOM = 68
RANK_MAX = 15
LINE_COLORS = [
    "#163a5f",
    "#9d2c00",
    "#46734b",
    "#6b4f9f",
    "#b26a00",
    "#3d6f8e",
    "#8d3d57",
    "#557a95",
    "#6a7f3a",
    "#9a5d56",
]
SVG_STYLE_BLOCK = """
    <style>
        .chart-bg { fill: #fffdfa; }
        .grid { stroke: #d9d3c9; stroke-width: 1; stroke-dasharray: 3 5; }
        .date-grid { stroke: #ebe5da; stroke-width: 1; }
        .axis-line { stroke: #98918a; stroke-width: 1.2; }
        .axis-label { fill: #615a52; font-size: 12px; font-family: system-ui, sans-serif; }
        .date-label { text-anchor: middle; }
        .rank-label { text-anchor: end; }
        .axis-title { fill: #36312d; font-size: 15px; font-family: system-ui, sans-serif; }
        .band-label { fill: #f3eee7; font-size: 16px; font-family: system-ui, sans-serif; font-weight: 600; text-anchor: middle; }
        .not-on-list-band { fill: #6b6865; opacity: 0.96; }
        .series-line { fill: none; stroke-width: 3.6; stroke-linecap: round; stroke-linejoin: round; opacity: 0.95; }
    </style>
"""


@dataclass
class AmazonCandidate:
    title: str
    author: str


@dataclass
class NytSeries:
    title: str
    author: str
    list_slug: str
    list_title: str
    ranks_by_date: Dict[str, int]
    page_dates: Set[str]
    best_rank: int
    amazon_title: str = ""
    amazon_match_score: float = 0.0

    @property
    def observed_weeks(self) -> int:
        return len(self.page_dates)

    @property
    def label(self) -> str:
        return f"{self.title} ({self.list_title})"


def read_rows(path: str | Path) -> List[Dict[str, str]]:
    with Path(path).open("r", newline="", encoding="utf-8") as handle:
        return list(csv.DictReader(handle))


def parse_int(value: str) -> Optional[int]:
    text = str(value).strip().replace(",", "")
    if not text:
        return None
    try:
        return int(float(text))
    except ValueError:
        return None


def normalize_text(value: str) -> str:
    text = str(value or "").lower().replace("&", " and ")
    text = re.sub(r"\bby\b", " ", text)
    text = re.sub(r"[^a-z0-9]+", " ", text)
    return " ".join(text.split())


def title_variants(title: str) -> Set[str]:
    normalized = normalize_text(title)
    variants = {normalized}
    for separator in (":", "(", "["):
        head = title.split(separator, 1)[0]
        normalized_head = normalize_text(head)
        if len(normalized_head.split()) >= 2:
            variants.add(normalized_head)
    return {variant for variant in variants if variant}


def author_tokens(author: str) -> Set[str]:
    normalized = normalize_text(author)
    return {token for token in normalized.split() if len(token) >= 3 and token not in {"with", "and", "the"}}


def overlap_ratio(left: Set[str], right: Set[str]) -> float:
    if not left or not right:
        return 0.0
    return len(left & right) / min(len(left), len(right))


def match_score(nyt_title: str, nyt_author: str, amazon_title: str, amazon_author: str) -> float:
    nyt_variants = title_variants(nyt_title)
    amazon_variants = title_variants(amazon_title)
    if nyt_variants & amazon_variants:
        return 1.0

    nyt_full = max(nyt_variants, key=len, default="")
    amazon_full = max(amazon_variants, key=len, default="")
    if not nyt_full or not amazon_full:
        return 0.0

    sequence_ratio = similarity_ratio(nyt_full, amazon_full)
    nyt_tokens = set(nyt_full.split())
    amazon_tokens = set(amazon_full.split())
    token_ratio = overlap_ratio(nyt_tokens, amazon_tokens)
    author_overlap = bool(author_tokens(nyt_author) & author_tokens(amazon_author))

    if nyt_full in amazon_full and len(nyt_tokens) >= 2:
        return 0.94
    if amazon_full in nyt_full and len(amazon_tokens) >= 2:
        return 0.94
    if token_ratio == 1.0 and sequence_ratio >= 0.7:
        return 0.9
    if author_overlap and sequence_ratio >= 0.72 and token_ratio >= 0.5:
        return 0.84
    if sequence_ratio >= 0.86 and token_ratio >= 0.6:
        return 0.8
    return 0.0


def similarity_ratio(left: str, right: str) -> float:
    if not left or not right:
        return 0.0
    if left == right:
        return 1.0
    left_bigrams = {left[index:index + 2] for index in range(max(len(left) - 1, 1))}
    right_bigrams = {right[index:index + 2] for index in range(max(len(right) - 1, 1))}
    if not left_bigrams or not right_bigrams:
        return 0.0
    return (2 * len(left_bigrams & right_bigrams)) / (len(left_bigrams) + len(right_bigrams))


def load_nyt_series(paths: Sequence[str | Path]) -> Tuple[List[NytSeries], List[str]]:
    dedupe_keys: Set[Tuple[str, str, str, str, str]] = set()
    by_series: Dict[Tuple[str, str], NytSeries] = {}
    all_dates: Set[str] = set()

    for path in paths:
        for row in read_rows(path):
            page_date = str(row.get("page_date", "")).strip()
            title = str(row.get("title", "")).strip()
            list_slug = str(row.get("list_slug", "")).strip()
            rank = parse_int(row.get("rank", ""))
            if not page_date or not title or not list_slug or rank is None:
                continue

            dedupe_key = (
                page_date,
                list_slug,
                str(row.get("isbn13", "")).strip(),
                title,
                str(rank),
            )
            if dedupe_key in dedupe_keys:
                continue
            dedupe_keys.add(dedupe_key)
            all_dates.add(page_date)

            series_key = (title, list_slug)
            if series_key not in by_series:
                by_series[series_key] = NytSeries(
                    title=title,
                    author=str(row.get("author", "")).strip(),
                    list_slug=list_slug,
                    list_title=str(row.get("list_title", list_slug)).strip() or list_slug,
                    ranks_by_date={},
                    page_dates=set(),
                    best_rank=rank,
                )

            series = by_series[series_key]
            existing_rank = series.ranks_by_date.get(page_date)
            if existing_rank is None or rank < existing_rank:
                series.ranks_by_date[page_date] = rank
            series.page_dates.add(page_date)
            series.best_rank = min(series.best_rank, rank)

    sorted_dates = sorted(all_dates)
    sorted_series = sorted(by_series.values(), key=lambda item: (-item.observed_weeks, item.best_rank, item.title, item.list_slug))
    return sorted_series, sorted_dates


def load_amazon_candidates(path: str | Path) -> List[AmazonCandidate]:
    deduped: Dict[Tuple[str, str], AmazonCandidate] = {}
    for row in read_rows(path):
        title = str(row.get("title", "")).strip()
        author = str(row.get("author", "")).strip()
        if not title:
            continue
        key = (normalize_text(title), normalize_text(author))
        deduped[key] = AmazonCandidate(title=title, author=author)
    return list(deduped.values())


def classify_series(series_list: Sequence[NytSeries], amazon_candidates: Sequence[AmazonCandidate]) -> Tuple[List[NytSeries], List[NytSeries]]:
    matched: List[NytSeries] = []
    unmatched: List[NytSeries] = []
    for series in series_list:
        best_score = 0.0
        best_candidate: Optional[AmazonCandidate] = None
        for candidate in amazon_candidates:
            score = match_score(series.title, series.author, candidate.title, candidate.author)
            if score > best_score:
                best_score = score
                best_candidate = candidate
        if best_candidate and best_score >= 0.8:
            series.amazon_title = best_candidate.title
            series.amazon_match_score = best_score
            matched.append(series)
        else:
            unmatched.append(series)
    return matched, unmatched


def series_values(series: NytSeries, page_dates: Sequence[str]) -> List[int]:
    return [series.ranks_by_date.get(page_date, NOT_ON_LIST_RANK) for page_date in page_dates]


def rank_to_y(rank_value: int) -> float:
    inner_height = CHART_HEIGHT - CHART_TOP - CHART_BOTTOM
    y_min = CHART_TOP
    y_max = CHART_HEIGHT - CHART_BOTTOM
    return y_min + ((rank_value - 1) / (NOT_ON_LIST_RANK - 1)) * inner_height


def step_path(values: Sequence[int]) -> str:
    if not values:
        return ""
    inner_width = CHART_WIDTH - CHART_LEFT - CHART_RIGHT
    x_step = inner_width / max(len(values) - 1, 1)
    commands: List[str] = []
    last_x = CHART_LEFT
    last_y = rank_to_y(values[0])
    commands.append(f"M {last_x:.1f} {last_y:.1f}")
    for index, value in enumerate(values[1:], start=1):
        x_value = CHART_LEFT + (index * x_step)
        current_y = rank_to_y(value)
        commands.append(f"L {x_value:.1f} {last_y:.1f}")
        commands.append(f"L {x_value:.1f} {current_y:.1f}")
        last_x = x_value
        last_y = current_y
    return " ".join(commands)


def build_chart_svg(series_list: Sequence[NytSeries], page_dates: Sequence[str]) -> str:
    chart_bottom_y = CHART_HEIGHT - CHART_BOTTOM
    band_top = rank_to_y(RANK_MAX + 0.2)
    grid_lines = []
    for rank in range(1, RANK_MAX + 1):
        y_value = rank_to_y(rank)
        grid_lines.append(
            f'<line x1="{CHART_LEFT}" y1="{y_value:.1f}" x2="{CHART_WIDTH - CHART_RIGHT}" y2="{y_value:.1f}" class="grid" />'
            f'<text x="{CHART_LEFT - 12}" y="{y_value + 4:.1f}" class="axis-label rank-label">{rank}</text>'
        )

    inner_width = CHART_WIDTH - CHART_LEFT - CHART_RIGHT
    x_step = inner_width / max(len(page_dates) - 1, 1)
    date_labels = []
    for index, page_date in enumerate(page_dates):
        x_value = CHART_LEFT + (index * x_step)
        date_labels.append(f'<line x1="{x_value:.1f}" y1="{CHART_TOP}" x2="{x_value:.1f}" y2="{chart_bottom_y}" class="date-grid" />')
        date_labels.append(f'<text x="{x_value:.1f}" y="{CHART_HEIGHT - 20}" class="axis-label date-label">{html.escape(page_date)}</text>')

    paths = []
    for index, series in enumerate(series_list):
        values = series_values(series, page_dates)
        final_x = CHART_LEFT + ((len(page_dates) - 1) * x_step if page_dates else 0)
        final_y = rank_to_y(values[-1]) if values else rank_to_y(NOT_ON_LIST_RANK)
        color = LINE_COLORS[index % len(LINE_COLORS)]
        paths.append(
            f'<path d="{step_path(values)}" class="series-line" style="stroke:{color}" />'
            f'<circle cx="{final_x:.1f}" cy="{final_y:.1f}" r="4" style="fill:{color}" />'
        )

    return f'''
    <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 {CHART_WIDTH} {CHART_HEIGHT}" class="trajectory-chart" role="img" aria-label="NYT rank trajectory chart">
      {SVG_STYLE_BLOCK}
      <rect x="0" y="0" width="{CHART_WIDTH}" height="{CHART_HEIGHT}" class="chart-bg" />
      <rect x="{CHART_LEFT}" y="{band_top:.1f}" width="{CHART_WIDTH - CHART_LEFT - CHART_RIGHT}" height="{chart_bottom_y - band_top:.1f}" class="not-on-list-band" />
      {''.join(grid_lines)}
      {''.join(date_labels)}
      <line x1="{CHART_LEFT}" y1="{chart_bottom_y}" x2="{CHART_WIDTH - CHART_RIGHT}" y2="{chart_bottom_y}" class="axis-line" />
      <text x="20" y="{(CHART_TOP + chart_bottom_y) / 2:.1f}" class="axis-title axis-title-y" transform="rotate(-90, 20, {(CHART_TOP + chart_bottom_y) / 2:.1f})">Position on NYT list</text>
      <text x="{CHART_WIDTH / 2:.1f}" y="{band_top + 28:.1f}" class="band-label">Not on list</text>
      {''.join(paths)}
    </svg>
    '''


def render_series_table(series_list: Sequence[NytSeries], show_amazon_column: bool) -> str:
    header_cells = ["Book", "NYT list", "Observed weeks", "Best rank"]
    if show_amazon_column:
        header_cells.extend(["Matched Amazon title", "Match score"])

    rows = []
    for series in series_list:
        cells = [
            html.escape(series.title),
            html.escape(series.list_title),
            str(series.observed_weeks),
            str(series.best_rank),
        ]
        if show_amazon_column:
            cells.extend([
                html.escape(series.amazon_title or "Not observed"),
                f"{series.amazon_match_score:.2f}" if series.amazon_match_score else "",
            ])
        rows.append("<tr>" + "".join(f"<td>{cell}</td>" for cell in cells) + "</tr>")

    return (
        "<table><thead><tr>"
        + "".join(f"<th>{html.escape(cell)}</th>" for cell in header_cells)
        + "</tr></thead><tbody>"
        + "".join(rows)
        + "</tbody></table>"
    )


def render_section(title: str, description: str, series_list: Sequence[NytSeries], page_dates: Sequence[str], show_amazon_column: bool) -> str:
    if not series_list:
        return f"<section class=\"panel\"><h2>{html.escape(title)}</h2><p>{html.escape(description)}</p><p>No series available with the current inputs.</p></section>"
    return f'''
    <section class="panel">
      <div class="panel-header">
        <div>
          <h2>{html.escape(title)}</h2>
          <p>{html.escape(description)}</p>
        </div>
        <div class="panel-metric">
          <span class="metric-value">{len(series_list)}</span>
          <span class="metric-label">charted titles</span>
        </div>
      </div>
      {build_chart_svg(series_list, page_dates)}
      {render_series_table(series_list, show_amazon_column=show_amazon_column)}
    </section>
    '''


def render_book_card(series: NytSeries, page_dates: Sequence[str], card_title: str, include_amazon_note: bool) -> str:
    values = series_values(series, page_dates)
    visible_dates = [page_date for page_date in page_dates if page_date in series.ranks_by_date]
    summary_bits = [
        f"Observed weeks: {series.observed_weeks}",
        f"Best NYT rank: {series.best_rank}",
        f"List: {series.list_title}",
    ]
    if include_amazon_note and series.amazon_title:
        summary_bits.append(f"Amazon match: {series.amazon_title}")

    return f'''
    <article class="book-card">
      <div class="book-card-header">
        <h3>{html.escape(card_title)}</h3>
        <p>{html.escape(' | '.join(summary_bits))}</p>
      </div>
      {build_chart_svg([series], page_dates)}
      <p class="book-card-dates">Observed NYT page dates: {html.escape(', '.join(visible_dates) if visible_dates else 'None')}</p>
      <div class="book-card-rank-row">{''.join(f'<span class="rank-pill">{html.escape(page_date)}: {value if value < NOT_ON_LIST_RANK else "not on list"}</span>' for page_date, value in zip(page_dates, values))}</div>
    </article>
    '''


def render_book_card_section(title: str, description: str, series_list: Sequence[NytSeries], page_dates: Sequence[str], include_amazon_note: bool) -> str:
    if not series_list:
        return ""
    cards = []
    for series in series_list:
        cards.append(
            render_book_card(
                series,
                page_dates,
                card_title=f"How {series.title} ranked over time",
                include_amazon_note=include_amazon_note,
            )
        )
    return f'''
    <section class="panel panel-cards">
      <div class="panel-header">
        <div>
          <h2>{html.escape(title)}</h2>
          <p>{html.escape(description)}</p>
        </div>
      </div>
      <div class="book-card-list">
        {''.join(cards)}
      </div>
    </section>
    '''


def build_html(
    matched: Sequence[NytSeries],
    unmatched: Sequence[NytSeries],
    page_dates: Sequence[str],
    top_matched: int,
    top_unmatched: int,
    include_matched: bool = True,
    include_unmatched: bool = True,
    include_book_cards: bool = True,
    page_title: str = "NYT vs Amazon Comparison Charts",
    lead_text: str = "",
) -> str:
    matched_series = list(matched)[:top_matched] if include_matched else []
    unmatched_series = list(unmatched)[:top_unmatched] if include_unmatched else []
    lead = lead_text or (
        "These comparison charts use the NYT public archive and weekly snapshots you already collected, then split the trajectories into two groups: books also observed in the tracked Amazon bestseller categories and books not observed there. A missing NYT week is shown in the gray band at the bottom as ‘Not on list,’ matching the visual logic of the reference chart."
    )
    return f'''<!doctype html>
<html lang="en">
<head>
  <meta charset="utf-8">
    <title>{html.escape(page_title)}</title>
  <style>
    :root {{
      --paper: #f6f0e6;
      --panel: #fffdf9;
      --ink: #171717;
      --muted: #645f58;
      --line: #d7d0c4;
      --grid: #d9d3c9;
      --band: #6b6865;
      --band-text: #f3eee7;
      --shadow: rgba(34, 28, 19, 0.08);
    }}
    * {{ box-sizing: border-box; }}
    body {{ margin: 0; font-family: Georgia, "Times New Roman", serif; background: radial-gradient(circle at top, #fcfaf5 0%, var(--paper) 62%, #ece4d7 100%); color: var(--ink); }}
    main {{ max-width: 1240px; margin: 0 auto; padding: 32px 24px 48px; }}
    h1 {{ font-size: 48px; line-height: 0.95; margin: 0 0 10px; letter-spacing: -0.04em; }}
    .lead {{ max-width: 860px; color: var(--muted); font-size: 17px; line-height: 1.55; margin-bottom: 24px; }}
    .panel {{ background: var(--panel); border: 1px solid var(--line); border-radius: 22px; padding: 22px; margin-bottom: 22px; box-shadow: 0 18px 45px var(--shadow); }}
    .panel-header {{ display: flex; justify-content: space-between; gap: 18px; align-items: end; margin-bottom: 14px; }}
    h2 {{ margin: 0 0 6px; font-size: 26px; line-height: 1.05; }}
    p {{ margin: 0; color: var(--muted); line-height: 1.5; }}
    .panel-metric {{ min-width: 120px; padding: 12px 14px; border-radius: 16px; background: #f0eadf; text-align: center; }}
    .metric-value {{ display: block; font-size: 28px; font-weight: bold; color: #163a5f; }}
    .metric-label {{ display: block; color: var(--muted); font-size: 12px; text-transform: uppercase; letter-spacing: 0.08em; }}
    .trajectory-chart {{ width: 100%; height: auto; display: block; border-radius: 16px; overflow: hidden; margin: 8px 0 18px; }}
    .chart-bg {{ fill: #fffdfa; }}
    .grid {{ stroke: var(--grid); stroke-width: 1; stroke-dasharray: 3 5; }}
    .date-grid {{ stroke: #ebe5da; stroke-width: 1; }}
    .axis-line {{ stroke: #98918a; stroke-width: 1.2; }}
    .axis-label {{ fill: #615a52; font-size: 12px; font-family: system-ui, sans-serif; }}
    .date-label {{ text-anchor: middle; }}
    .rank-label {{ text-anchor: end; }}
    .axis-title {{ fill: #36312d; font-size: 15px; font-family: system-ui, sans-serif; }}
    .band-label {{ fill: var(--band-text); font-size: 16px; font-family: system-ui, sans-serif; font-weight: 600; text-anchor: middle; }}
    .not-on-list-band {{ fill: var(--band); opacity: 0.96; }}
    .series-line {{ fill: none; stroke-width: 3.6; stroke-linecap: round; stroke-linejoin: round; opacity: 0.95; }}
    table {{ width: 100%; border-collapse: collapse; font-family: system-ui, sans-serif; font-size: 14px; }}
    th, td {{ text-align: left; padding: 10px 12px; border-top: 1px solid #ece5d9; vertical-align: top; }}
    th {{ font-size: 12px; color: #6f675f; text-transform: uppercase; letter-spacing: 0.05em; }}
        .panel-cards {{ padding-top: 18px; }}
        .book-card-list {{ display: grid; grid-template-columns: 1fr; gap: 18px; }}
        .book-card {{ background: #fcfaf4; border: 1px solid #e5ddcf; border-radius: 18px; padding: 18px; }}
        .book-card-header h3 {{ margin: 0 0 6px; font-size: 24px; line-height: 1.02; letter-spacing: -0.03em; }}
        .book-card-header p {{ margin-bottom: 10px; }}
        .book-card-dates {{ font-family: system-ui, sans-serif; font-size: 13px; color: #6b645d; margin-top: -4px; }}
        .book-card-rank-row {{ display: flex; flex-wrap: wrap; gap: 8px; margin-top: 10px; }}
        .rank-pill {{ font-family: system-ui, sans-serif; font-size: 12px; color: #48423c; background: #efe7db; border-radius: 999px; padding: 6px 10px; }}
    @media (max-width: 920px) {{
      h1 {{ font-size: 36px; }}
      .panel-header {{ flex-direction: column; align-items: start; }}
      main {{ padding: 22px 14px 36px; }}
    }}
  </style>
</head>
<body>
  <main>
    <h1>{html.escape(page_title)}</h1>
    <p class="lead">{html.escape(lead)}</p>
    {render_section(
        title="NYT books also observed in tracked Amazon categories",
        description="Matched conservatively by normalized title, subtitle trimming, and author-aware fuzzy checks. These are the clearest cross-surface overlaps in the current tracked categories.",
        series_list=matched_series,
        page_dates=page_dates,
        show_amazon_column=True,
    ) if include_matched else ''}
    {render_section(
        title="NYT books not observed in tracked Amazon categories",
        description="These books appeared in the NYT inputs but were not found in the Amazon public history you are currently tracking. This is category-limited absence, not global Amazon absence.",
        series_list=unmatched_series,
        page_dates=page_dates,
        show_amazon_column=False,
    ) if include_unmatched else ''}
    {render_book_card_section(
        title="Standalone per-book charts for NYT books also observed on Amazon",
        description="Each card isolates one title so you can inspect the exact weekly NYT trajectory without line overlap.",
        series_list=matched_series,
        page_dates=page_dates,
        include_amazon_note=True,
    ) if include_book_cards and include_matched else ''}
    {render_book_card_section(
        title="Standalone per-book charts for NYT books not observed in tracked Amazon categories",
        description="These cards show the same NYT trajectories for the unmatched group, one book at a time.",
        series_list=unmatched_series,
        page_dates=page_dates,
        include_amazon_note=False,
    ) if include_book_cards and include_unmatched else ''}
  </main>
</body>
</html>
'''


def sibling_output_path(output_path: Path, suffix: str) -> Path:
    return output_path.with_name(f"{output_path.stem}_{suffix}{output_path.suffix}")


def sibling_output_dir(output_path: Path, suffix: str) -> Path:
    return output_path.with_name(f"{output_path.stem}_{suffix}")


def write_html(path: Path, content: str) -> Path:
    ensure_directory(path.parent)
    path.write_text(content, encoding="utf-8")
    return path


def write_text(path: Path, content: str) -> Path:
    ensure_directory(path.parent)
    path.write_text(content, encoding="utf-8")
    return path


def slugify(value: str) -> str:
    slug = normalize_text(value).replace(" ", "_")
    return slug[:80] or "untitled"


def choose_paper_series(series_list: Sequence[NytSeries], explicit_titles: Sequence[str], count: int) -> List[NytSeries]:
    if explicit_titles:
        desired = {normalize_text(title) for title in explicit_titles}
        selected = [series for series in series_list if normalize_text(series.title) in desired]
        return selected[:count] if count > 0 else selected
    return list(series_list)[:count]


def export_svg_figures(
    figure_dir: Path,
    matched_series: Sequence[NytSeries],
    unmatched_series: Sequence[NytSeries],
    page_dates: Sequence[str],
) -> List[Path]:
    svg_paths: List[Path] = []
    panels_dir = figure_dir / "panels"
    matched_books_dir = figure_dir / "matched_books"
    unmatched_books_dir = figure_dir / "unmatched_books"

    panel_specs = [
        (panels_dir / "matched_panel.svg", matched_series),
        (panels_dir / "unmatched_panel.svg", unmatched_series),
    ]
    for path, series_list in panel_specs:
        if not series_list:
            continue
        svg_paths.append(write_text(path, build_chart_svg(series_list, page_dates)))

    for index, series in enumerate(matched_series, start=1):
        path = matched_books_dir / f"{index:02d}_{slugify(series.title)}.svg"
        svg_paths.append(write_text(path, build_chart_svg([series], page_dates)))

    for index, series in enumerate(unmatched_series, start=1):
        path = unmatched_books_dir / f"{index:02d}_{slugify(series.title)}.svg"
        svg_paths.append(write_text(path, build_chart_svg([series], page_dates)))

    return svg_paths


def export_paper_ready_figures(
    figure_dir: Path,
    paper_series: Sequence[NytSeries],
    page_dates: Sequence[str],
) -> List[Path]:
    svg_paths: List[Path] = []
    panels_dir = figure_dir / "panels"
    matched_books_dir = figure_dir / "matched_books"
    if paper_series:
        svg_paths.append(write_text(panels_dir / "paper_matched_panel.svg", build_chart_svg(paper_series, page_dates)))
    for index, series in enumerate(paper_series, start=1):
        path = matched_books_dir / f"{index:02d}_{slugify(series.title)}.svg"
        svg_paths.append(write_text(path, build_chart_svg([series], page_dates)))
    return svg_paths


def export_png_figures(svg_paths: Sequence[Path]) -> List[Path]:
    from playwright.sync_api import sync_playwright

    png_paths: List[Path] = []
    with sync_playwright() as playwright:
        browser = playwright.chromium.launch()
        page = browser.new_page(viewport={"width": CHART_WIDTH + 32, "height": CHART_HEIGHT + 32}, device_scale_factor=2)
        for svg_path in svg_paths:
            page.goto(svg_path.resolve().as_uri())
            page.locator("svg").first.screenshot(path=str(svg_path.with_suffix(".png")))
            png_paths.append(svg_path.with_suffix(".png"))
        browser.close()
    return png_paths


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser(description="Render NYT-style comparison charts for NYT titles with and without Amazon-category overlap.")
    parser.add_argument(
        "--nyt-inputs",
        nargs="+",
        default=[
            "validation_data/nyt_public_hardcover_nonfiction_archive_2026Q2.csv",
            "validation_data/nyt_public_weekly_history.csv",
        ],
        help="NYT public history CSV inputs.",
    )
    parser.add_argument("--amazon-input", default="validation_data/amazon_public_history.csv", help="Amazon public history CSV input.")
    parser.add_argument("--output", default="validation_results/nyt_amazon_comparison_charts.html", help="Output HTML path.")
    parser.add_argument("--matched-output", default="", help="Optional matched-only HTML path. Defaults beside --output.")
    parser.add_argument("--unmatched-output", default="", help="Optional unmatched-only HTML path. Defaults beside --output.")
    parser.add_argument("--figures-dir", default="", help="Optional directory for standalone SVG and PNG chart exports. Defaults beside --output.")
    parser.add_argument("--paper-html-output", default="", help="Optional paper-ready matched-subset HTML path. Defaults beside --output.")
    parser.add_argument("--paper-figures-dir", default="", help="Optional paper-ready SVG and PNG figure directory. Defaults beside --output.")
    parser.add_argument("--paper-top-matched", type=int, default=6, help="Number of matched titles to include in the paper-ready subset export.")
    parser.add_argument("--paper-titles", nargs="*", default=[], help="Optional explicit matched NYT titles to use for the paper-ready subset export.")
    parser.add_argument("--top-matched", type=int, default=12, help="Number of matched NYT series to include.")
    parser.add_argument("--top-unmatched", type=int, default=12, help="Number of unmatched NYT series to include.")
    return parser.parse_args()


def main() -> int:
    args = parse_args()
    nyt_series, page_dates = load_nyt_series(args.nyt_inputs)
    amazon_candidates = load_amazon_candidates(args.amazon_input)
    matched, unmatched = classify_series(nyt_series, amazon_candidates)

    output_path = Path(args.output)
    matched_output_path = Path(args.matched_output) if args.matched_output else sibling_output_path(output_path, "matched_only")
    unmatched_output_path = Path(args.unmatched_output) if args.unmatched_output else sibling_output_path(output_path, "unmatched_only")
    figures_dir = Path(args.figures_dir) if args.figures_dir else sibling_output_dir(output_path, "figures")
    paper_html_output_path = Path(args.paper_html_output) if args.paper_html_output else sibling_output_path(output_path, "paper_ready")
    paper_figures_dir = Path(args.paper_figures_dir) if args.paper_figures_dir else sibling_output_dir(output_path, "paper_ready_figures")

    matched_series = list(matched)[: args.top_matched]
    unmatched_series = list(unmatched)[: args.top_unmatched]
    paper_series = choose_paper_series(matched_series, args.paper_titles, args.paper_top_matched)

    content = build_html(
        matched_series,
        unmatched_series,
        page_dates,
        top_matched=args.top_matched,
        top_unmatched=args.top_unmatched,
        page_title="NYT Trajectories vs Amazon Presence",
    )
    matched_content = build_html(
        matched_series,
        unmatched_series,
        page_dates,
        top_matched=args.top_matched,
        top_unmatched=args.top_unmatched,
        include_unmatched=False,
        page_title="NYT Titles Also Observed In Tracked Amazon Categories",
        lead_text="This page isolates the NYT titles that also appear in the tracked Amazon bestseller categories, shown first as a multi-line comparison and then as standalone per-book charts.",
    )
    unmatched_content = build_html(
        matched_series,
        unmatched_series,
        page_dates,
        top_matched=args.top_matched,
        top_unmatched=args.top_unmatched,
        include_matched=False,
        page_title="NYT Titles Not Observed In Tracked Amazon Categories",
        lead_text="This page isolates NYT titles that were not observed in the currently tracked Amazon bestseller categories, shown as a multi-line comparison and standalone per-book charts.",
    )
    paper_content = build_html(
        paper_series,
        [],
        page_dates,
        top_matched=len(paper_series),
        top_unmatched=0,
        include_unmatched=False,
        page_title="Paper-Ready Matched NYT Trajectories",
        lead_text="This page narrows the export to a smaller matched subset intended for paper figures and faster review.",
    )

    write_html(output_path, content)
    write_html(matched_output_path, matched_content)
    write_html(unmatched_output_path, unmatched_content)
    write_html(paper_html_output_path, paper_content)
    svg_paths = export_svg_figures(figures_dir, matched_series, unmatched_series, page_dates)
    png_paths = export_png_figures(svg_paths)
    paper_svg_paths = export_paper_ready_figures(paper_figures_dir, paper_series, page_dates)
    paper_png_paths = export_png_figures(paper_svg_paths)
    print(
        f"Wrote NYT/Amazon comparison charts to {output_path}, {matched_output_path}, and {unmatched_output_path} "
        f"plus paper-ready HTML {paper_html_output_path}, {len(svg_paths)} SVG files and {len(png_paths)} PNG files in {figures_dir}, "
        f"and {len(paper_svg_paths)} paper-ready SVG files and {len(paper_png_paths)} paper-ready PNG files in {paper_figures_dir} "
        f"({len(matched)} matched series, {len(unmatched)} unmatched series, {len(page_dates)} page dates)"
    )
    return 0


if __name__ == "__main__":
    raise SystemExit(main())